dismutase mutants
by
Ming Sze Tong
A thesis
presented to the University of Waterloo
in fulfilment of the
thesis requirement for the degree of
Master of Science
in
Chemistry
Authors Declaration
I hereby declare that I am the sole author of this thesis. This is a true copy of the thesis, including
any required final revisions, as accepted by my examiners.
I understand that my thesis may be made electronically available to the public.
ii
Abstract
Amyotrophic lateral sclerosis (ALS) is the most common motor neuron disease and is
characterized by progressive paralysis leading to death, typically, within 3-5 years of the onset of
symptoms. The majority of ALS cases are sporadic with no known causative agent; however, 510% of ALS cases are genetically inherited and termed familial ALS (fALS). Approximately,
15-20% of these fALS cases have been linked to mutations in the gene encoding human Cu/Zn
superoxide dismutase (hSOD). To date, over 140 hSOD mutations have been discovered. The
mechanisms by which mutant hSOD confers toxicity in fALS patients are still unknown.
However, there is growing evidence that ALS is a type of protein conformational disease
whereby cell damage or death is caused by the accumulation of protein aggregates in the cell. It
is hypothesized that mutations destabilise hSOD and increase its propensity to aggregate. There
is some controversy as to which hSOD species contributes to aggregation. Many believe that
only apo or mismetallated forms of hSOD are able to aggregate. Due to the abundance of fully
metallated or holo hSOD in the cell, we hypothesize that holo hSOD can also lead to
aggregation. Holo dimer interface mutants A4S, A4T and I113T as well as G41D were found to
be destabilized compared to holo pseudo wildtype (pWT) while zinc binding mutant H80R was
shown to form fragments via an unknown mechanism. Holo dimer interface mutants A4S and
A4T were also shown to have an increased propensity to aggregate compared to pWT, which
correlates to their decreased stability as well a short disease durations.
iii
Acknowledgements
First and foremost, I would like to thank Dr. E. M. Meiering for giving me the opportunity to
pursue the research presented in this thesis. I would also like to thank her for all her guidance
and insight, both academically and on life.
I would also like to thank the members of the Meiering laboratory, past and present, for all their
help and support. I would especially like to thank Gracie Hwang, Dr. Jessica Rumfeldt, Helen
Stubbs and Kenrick Vassall.
Last but not least, I would like to thank my parents, Tong Chap Thy and Teoh Teik Gek, for
supporting me throughout my life, through many years of formal education and for all the
sacrifices they have made for me to be where I am today. And to my partner, Andi, I thank you
for your love and support, and your unending faith in me.
iv
Dedication
For papa & mama
and katze
Table of Contents
Authors Declaration ....................................................................................................................... ii
Abstract .......................................................................................................................................... iii
Acknowledgements ........................................................................................................................ iv
Dedication ....................................................................................................................................... v
Table of Contents ........................................................................................................................... vi
List of Tables .................................................................................................................................. x
List of Figures ................................................................................................................................ xi
List of Abbreviations ................................................................................................................... xiii
Chapter 1 General introduction ....................................................................................................... 1
1.1
1.2
1.3
1.5
1.6
Introduction ................................................................................................................. 12
Methods....................................................................................................................... 16
2.3
Results ......................................................................................................................... 20
2.3.1 Obtaining recombinant fALS-associated hSOD plasmids using PCR-mediated sitedirected mutagenesis ................................................................................................... 20
2.3.2 Expression of recombinant hSOD .............................................................................. 21
2.3.3 Confirmation of recombinant hSOD mass .................................................................. 21
2.4
Discussion ................................................................................................................... 22
Introduction ................................................................................................................. 24
Methods....................................................................................................................... 27
Results ......................................................................................................................... 32
Discussion ................................................................................................................... 51
3.4.1 Dimer interface mutants and G41D are destabilized compared to pWT in the holo
state ............................................................................................................................. 51
3.4.2 Dimer interface mutants have decreased dimer stability ............................................ 52
3.4.3 H80R ........................................................................................................................... 54
3.4.3 Correlations with disease .............................................................................................. 55
Chapter 4 In vitro aggregation of holo pWT and dimer interace mutants .................................... 57
4.1
Introduction ................................................................................................................. 57
Methods....................................................................................................................... 62
Acknowledgements ..................................................................................................... 64
4.4
Results ......................................................................................................................... 64
4.4.3 Holo hSOD aggregation proceeds via a secondary nucleation mechanism ................ 69
4.5
Discussion ................................................................................................................... 71
4.5.1 Aggregates arise from holo hSOD dimer interface mutants A4S and A4T at
physiologically relevant pH and temperature ............................................................. 71
4.5.2 Holo hSOD dimer interface mutants A4S and A4T have shorter lag times compared
to pWT ........................................................................................................................ 73
4.5.3 Multiple pathways of aggregation .............................................................................. 75
Chapter 5 Summary and future work ............................................................................................ 77
5.1
5.2
ix
List of Tables
Table 1.1 The amino acid sequence of hSOD along with fALS-associated substitution
mutations .........................................................................................................................8
Table 2.1 Summary of available patient data ................................................................................15
Table 2.2 Primer sequences for A4S, A4T, G37R, G41D, H48R, H80R, N86D, I112T and
V148G/I with %GC and tm. ...........................................................................................17
Table 2.3 PCR components for master mix...................................................................................18
Table 2.4 Conditions for thermocycling ....................................................................................... 18
Table 2.5 Summary of recombinant hSOD masses. ......................................................................22
Table 3.1 Fitted dimer 2-state parameters for holo dimer interface mutants ................................37
x
Table 3.2 Dissociation constants, Kd, and molecularity for pWT and dimer interface mutants ...38
Table 3.3 Thermodynamic parameters of holo fALS-associated mutants A4S, A4T, I113T and
G41D compared to pWT. ..............................................................................................41
Table 4.1 Summary of holo hSOD aggregation fitted parameters to Equation 4.5.......................69
List of Figures
Figure 1.1 A ribbon representation of crystal structure (A) and schematic (B) of the hSOD
dimer. .............................................................................................................................5
Figure 2.1 A schematic of the pPHSOD1ASlacIq vector ..............................................................13
Figure 2.2 Mutation sites in hSOD associated to ALS..................................................................14
Figure 2.3 PCR products from PCR-mediated site directed mutagenesis of ALS-associated
hSOD mutants ..............................................................................................................20
Figure 2.4 SDS-PAGE of recombinant hSOD mutants obtained from expression .......................21
Figure 3.1 The specific heat capacity function of a globular protein. ...........................................25
Figure 3.2 Dimer 2-state and monomer 2-state fits of holo dimer interface mutants ...................34
Figure 3.3 Protein concentration dependence of holo pWT and holo dimer interface mutants ....35
Figure 3.4 Protein concentration dependence of
.....................................................................36
Figure 3.5 ALS-associated mutants are destabilized relative to pWT in the holo state ................40
Figure 3.6 Change in H80R species distribution over time ..........................................................43
Figure 3.7 SDS-PAGE of H80R OS incubated under different conditions ..................................45
Figure 3.8 Mass spectrum of fragmented H80R in non-reducing (A) and reducing (B) conditions.
......................................................................................................................................46
Figure 3.9 SDS-PAGE of H80R grown and heat treated with different conditions .....................48
Figure 3.10 DSC thermograms of as isolated (d and e) and apo (a-c) H80R ................................50
Figure 3.11 Site of cleavage in H80R ............................................................................................55
Figure 4.1 Autocorrelation functions of a large and small particle ...............................................59
Figure 4.2 A graphical representation of the sigmoidal increase in mean light scattering intensity
upon aggregate formation ............................................................................................61
xi
xii
List of Abbreviations
AD ..................................................................................................................... Alzheimers disease
ALS ......................................................................................................amyotrophic lateral sclerosis
apo hSOD .......................................... metal free form of human copper zinc superoxide dismutase
BSA ................................................................................................................ bovine serum albumin
................................................................................................................... specific heat capacity
DLS ............................................................................................................. dynamic light scattering
DSC ............................................................................................... differential scanning calorimetry
EDTA .............................................................................................. ethylenediaminetetraacetic acid
fALS ....................................................................................... familial amyotrophic lateral sclerosis
HD .................................................................................................................... Huntingtons disease
HEPES ............................................................4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
xiii
Holo hSOD............................... fully metallated form of human copper zinc superoxide dismutase
hSOD................................................................................human copper zinc superoxide dismutase
IPTG................................................................................... isopropyl beta-D-thiogalactopyranoside
ICP-AES .............................................. inductively-coupled plasma atomic emission spectroscopy
ITC ...................................................................................................isothermal titration calorimetry
kcal ................................................................................................................................... kilocalorie
..................................................................................................................... dissociation constant
LB ..................................................................................................................................... luria broth
MBR................................................................................................................. metal binding region
OS ...............................................................................................................................osmotic shock
PCD .................................................................................................. protein conformational disease
PD ...................................................................................................................... Parkinsons disease
xiii
conversion of a specific protein or proteins from a native fold to a non-native fold that eventually
leads to the formation cytotoxic oligomers or aggregates (1). As many as forty types of PCDs
have been discovered to date (2,3). The different proteins involved in PCDs do not share any
similarities in primary sequence or native structure; however, they share a vital characteristic
whereby the proteins involved have the ability to adopt certain conformational changes that
enable the formation of insoluble aggregates (3). Protein folding is a complicated process that
includes many opportunities for side reactions that can lead to aggregate-formation (4). The
majority of cases of PCDs are sporadic with no known causative agents. However, a small
1
portion of PCDs are associated with mutations in their associated proteins. These familial forms
of PCDs may provide some insight to the pathogenesis of the diseases. The mutations that occur
in the disease associated proteins tend to be dominantly inherited and are usually associated with
earlier disease onset (5), implying that these mutations possess some characteristics that confer
increased toxicity.
Like many neurodegenerative diseases, such as Alzheimers disease (AD), Parkinsons
disease (PD), Huntingtons disease (HD) and transmissible spongiform encephalopathies (TSE),
amyotrophic lateral sclerosis (ALS) has been proposed to be a PCD (2). However, unlike these
diseases, ALS is not an amyloid disease. Amyloid diseases are characterised by positive Congo
red staining with green-gold birefringence when viewed under polarized light (6-8) and amyloid
deposits with rigid unbranched fibrils with diameters of 5-13 nm and lengths of 0.1-1.6 m
1
(7,9,10). On the other hand, deposits found in ALS patients do not bind Congo red (11) and
contain filamentous as well as granule-coated fibrils (12-14).
1.2
prevalence of 1.5-2.5 in 100,000 individuals per year (15). To date, no cure has been found for
ALS and ALS treatment has only been able to slightly delay disease progression (16,17).
Amyotrophic lateral sclerosis is also referred to as Charcots disease, after Jean-Martin Charcot
who characterized the disease over 130 years ago, and more commonly as Lou Gehrigs disease,
after a famous American baseball player who was stricken by the disease in the late 1930s
(18,19). Amyotrophic lateral sclerosis is a late onset disease with an average onset age of 46
years (range 24-72) and is usually fatal within 3-5 years (range 0.3-20 years) of the onset of
symptoms (20). Early symptoms of ALS usually are weakness in the arms and legs which
progressively degenerates, inevitably leading to paralysis and death, usually due to respiratory
failure. There are also cases of bulbar onset which then extends outwards. The disease selectively
targets the motor neurons in the spinal cord, brain stem and cortex which usually leave the
patients intellect unaffected.
The majority of ALS cases have no known causative agents and are termed sporadic ALS
(sALS). However, a small percentage of ALS cases, about 5-10%, are genetically inherited and
termed familial ALS (fALS). In 1993, Rosen et al. discovered a genetic link between a subset of
fALS cases and mutations in the sod1 gene (21), which codes for human Cu/Zn superoxide
dismutase (hSOD). Since then, over 140 fALS linked hSOD mutations, representing
approximately 20% of all fALS cases (19), have been discovered (database containing hSOD
2
harmful effects of superoxide, a reactive oxygen species, and is ubiquitously found in all aerobic
organisms (24). In eukaryotes, it is found in the cytosol, nucleus, peroxisomes and mitochondrial
intermembrane space (19). The human Cu/Zn superoxide dismutase (hSOD) is encoded in the
sod1 gene located on the long arm of chromosome 21, more specifically at 21q22 (25,26). It is a
single copy gene spanning 11 kilobases and contains five exons and four introns.
1.3.1 Structure and function
Human Cu/Zn superoxide dismutase is a very stable homodimeric enzyme. The
homodimer is 32 kDa in size and each monomer consists of 153 amino acids. Each monomer is
composed of an 8-stranded Greek key -barrel. The eight -strands are antiparallel and are joined
by loops (27). Each monomer also contains an active site that preferentially and very tighly binds
one Cu2+ and one Zn2+ (Figure 1.1) (28). The active site on the surface of the -barrel and is
enclosed by loop 4, which is the zinc binding loop that extends into the dimer interface, and loop
7, which is the electrostatic loop. Each hSOD monomer has four cysteines; Cys-57 and Cys-146
3
form a conserved intramolecular disulfide bond while Cys-6 and Cys-111 are free cysteines (28).
The structure of the enzyme as well as the active site residues are highly conserved (19,27).
Figure 1.1 is a representation of the crystal structure of pseudo wildtype (pWT) hSOD (Section
2.1.1) with the intramolecular disulfide bonds and metals shown along with a schematic
illustrating the Greek key motif of hSOD. pWT is a C6S/C111A double mutant that removes the
2 free cysteines present in hSOD.
Figure 1.1 A ribbon representation of crystal structure (A) and schematic (B) of the hSOD
dimer. (A) Prepared using MolMol (29) and PDB coordinates 1SOS (30). The eight -strands
are in blue while the three -helices are in pink and yellow. Each monomer contains a Zn2+ and a
Cu2+ ion depicted as black and orange spheres, respectively. In both figures, the intramolecular
disulfide bond is shown in red while the zinc and electrostatic loops are in cyan and yellow,
respectively.
5
Both metals in hSOD are important for the structural stability of the active site. The
copper ion is an essential cofactor in hSOD catalysis and is cyclically oxidized and reduced
during catalysis. Human Cu/Zn superoxide dismutase catalyzes the dismutation of two
superoxide molecules to dioxygen and hydrogen peroxide (Reactions 1-3, below) (19). The
catalysis begins with the reduction of the cupric ion by the first molecule of superoxide to
produce molecular oxygen (Reaction 1) (19). The second superoxide molecule then reoxidizes
the cuprous ion to produce hydrogen peroxide (Reaction 2) (19). Superoxide dismutase is among
the fastest enzymes known with a rate constant of ~2 x 109 M-1s-1 (31).
O2 - + Cu2+ZnSOD O2 + Cu+ZnSOD
Reaction 1
Reaction 2
Overall reaction
2O2 - + 2 H+ O2 + H2O2
1.4
Reaction 3
substitutions at one of 74 different residues (shown in Table 1.1) (mutation list obtained from an
online database http://alsod.iop.kcl.ac.uk/Als/index.aspx). Other sod1 mutations include
deletions or insertions, frameshift mutations and truncations (not indicated in Table 1.1) (19,32).
All known sod1 mutations are dominantly inherited except for D90A which exhibits variable
inheritance (33-35). Although mutations are scattered throughout the protein, some of the
mutations can be categorized into two distinct groups: metal binding region (MBR) mutants and
dimer interface mutants. Metal binding region mutants are mutations that occur at the metal
coordinating residues as well as residues in the electrostatic and zinc loops. These affect the
6
metal binding ability of the protein, and are expected to result in lower enzymatic activity (19).
Dimer interface mutants, on the other hand, are mutations that occur in and around the dimer
interface and may affect the stability of the protein. To date, no clear trends have been identified
among the different types of hSOD mutations. However, a recent computational study has
proposed that nearly all mutations, regardless of their position, decrease dimer stability and/or
dissociation (36).
Table 1.1 The amino acid sequence of hSOD along with fALS-associated substitution
mutations. The mutations are listed in red. Boxes shaded blue indicate residues in -strands. The
circles above specific residues denote metal coordinating residues where the copper and zinc are
represented by yellow and black circles, respectively (mutation list obtained from an online
database http://alsod.iop.kcl.ac.uk/Als/index.aspx).
1
10
11
12
13
14
15
16
17
18
19
20
ST
V
F
G
Q
V
R
V
G
M
A
S
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
R
V
K
G
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A
F
Q
R
D
S
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
C
S
V
Y
81
82
83
84
85
86
87
88
89
90
91
92
93*
94
95
96
97
98
99
100
F
V
R
S
D I
KS
A
M
T
V
A
V
T
V
N
V
L
M
K
G
AC
DR
101
102
103
104
105
106
107
108
109
110
111
F
V
GH
NY
112
113
114
115
116
117
118
119
120
M
T
F
T
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
G
V
DH
K
141
142
143
144
145
146
147
148
149
150
151
152
153
F
S
G
T
D
R
G
I
S
T
* G93 has an additional two mutations not listed in the table; G93S and G93V
1.5
that hSOD toxicity was due to a loss or decrease in enzymatic activity (23). However, many
studies indicate that hSOD toxicity is caused by a toxic gain-of-function as opposed to a loss of
function. For example, transgenic mice expressing fALS-associated hSOD mutants show ALSlike symptoms despite having elevated hSOD activity (37). More compelling evidence against
the loss of function theory is the fact that hSOD knockout transgenic mice do not develop motor
neuron disease and appear to have a normal lifespan (38). Other studies showed that transgenic
mice coexpressing endogenous mouse SOD with fALS-associated hSOD mutants still develop
ALS-like symptoms and overexpressing wild-type hSOD with mutant hSOD either did not
alleviate the symptoms or accelerated disease progression (39), suggesting that cytotoxicity is
due to some intrinsic property in the hSOD mutants.
Despite the strong consensus for the toxic gain-of-function theory, the specific
mechanism of pathogenesis has yet to be elucidated. The two main gain-of-toxic function
hypotheses, which are not mutually exclusive, are: 1) oxidative damage hypothesis and 2)
protein aggregation hypothesis. The oxidative damage hypothesis proposes that ALS-associated
mutations reduce the specificity of the active site, increasing the probability of harmful, nonnative chemistry and the generation of free radicals. In the presence of hydrogen peroxide,
mutant hSOD have been shown to generate hydroxyl radicals that can lead to oxidative damage
and deactivation of the enzyme (40-42). Additionally, G93A mouse model studies found
increased levels of oxygen radicals and oxidative damage in the spinal cords of the mice (43,44).
HSOD can also react with peroxynitrite to produce a nitronium-like intermediate that can then
nitrate tyrosine residues leading to protein damage (45). Elevated levels of nitrotyrosine have
9
been found in motor neurons of sALS and fALS patients (46) as well as ALS mice models
(47,48). Despite the supporting evidence for the oxidative damage hypothesis, it is unlikely that
it is the common factor causing ALS. This is due to the fact that these mechanisms require the
presence of the active site copper and metal binding mutants that have decreased/no copper
binding still lead to the disease. Additionally, a mice model with all four copper binding residues
mutated, eliminating copper binding, nevertheless develope typical disease symptoms (49).
The protein aggregation hypothesis proposes that fALS-associated mutations increase
hSODs propensity to misfold and aggregate, and it is the aggregates that confer hSODs
cytotoxic properties. Indeed, protein misfolding and aggregation appear to be associated with
the pathogenesis of many neurodegenerative diseases such as Alzheimers disease (AD),
Parkinsons disease (PD) and Huntingtons disease (HD) (19). Proteinaceous inclusions from
some sALS and fALS patients (50-52), transgenic mice and cell culture models of ALS have all
been shown to be strongly immunoreactive to hSOD antibodies (37,39,53). In cell lines where
protein folding chaperones are co-expressed with mutant hSOD, aggregate formation is reduced
while cell viability increases (54), implicating that aggregates play a role in cell death. It is likely
that fALS-associated hSOD mutant pathogenesis is not exclusively caused by a single
mechanism; however, hSOD aggregation clearly plays a role in fALS pathology.
1.6
mutations increase the propensity of hSOD to aggregate. This thesis presents biophysical
analyses of fALS-associated dimer interface mutants A4S, A4T and I113T in terms of their
thermal stability and propensity to aggregate as well as preliminary stability data on two other
10
mutants, G41D and H80R. Chapter 2 describes the introduction of various structurally and
chemically diverse ALS-associated hSOD mutations to the pseudo wildtype (pWT) hSOD
background using polymerase chain reaction-site directed mutagenesis. Very little biophysical
data is available for many of these mutants and this has served to increase the catalogue of
available mutants for analysis. In Chapter 3, the thermal stability of fALS-associated mutants
A4T, A4S, I113T and G41D in the fully metallated (holo) state is determined using differential
scanning calorimetry (DSC). In addition, the mechanism as well as changes in enthalpy, entropy
and overall free energy of unfolding are analyzed. Preliminary characterization of the metal
binding mutant H80R is also included in this chapter. Chapter 4 focuses on the in vitro
aggregation mechanism of the holo form of the dimer interface mutants A4S and A4T.
Differences in aggregation propensity of the mutants compared to pWT are considered together
with thermal stability and epidemiological data to identify any correlations. The overall objective
is to elucidate the factors that confer toxicity to ALS-associated hSOD mutants, such as specific
characteristics of the mutants that promote aggregation, and to contribute to the collective
knowledge available on the disease, which may one day aid in the discovery of a cure.
11
Introduction
12
Figure 2.1 A schematic of the pPHSOD1ASlacIq vector. The plasmid contains the gene for
pWT (57) connected to a leader sequence from the SOD gene of Photobacterium leiognathi (58)
as well as genes encoding ampicillin resistance (ampr), the lac repressor (lacIq) and tacI
promoter (tacI) (59) (adapted from (60)).
2.1.3 Mutants of interest
To date, over 140 point mutations in hSOD have been associated with fALS. These
mutations are found throughout the protein, illustrated in Table1.1. Nearly all reported mutations
are dominantly inherited, with the exception of D90A which is recessively inherited (53) and a
few cases expressing varying degrees of penetrance (61,62). Clearly, it is important to study
these mutants and determine the properties they possess that that lead to ALS. The mutations
studied in this thesis are structurally (Figure 2.2) and chemically diverse which include dimer
13
interface mutations (A4S, A4T, I112T, V148G, V148I), metal binding region mutations (H48R
copper binding site mutant, H80R zinc binding site mutant) as well as mutations in the beta
barrel (G37R, G41D, N86D), which also charge mutations. Table 2.1 is a summary of available
patient data for the aforementioned mutations.
Figure 2.2 Mutation sites in hSOD associated to ALS that were studied in this thesis. Prepared
using MolMol (29) and PDB coordinates 1SOS (30). Each monomer contains a Zn2+ and a Cu2+
ion depicted as black and orange spheres, respectively. Ther dimer interface, MBR and charge
mutants are labelled in red, green and magenta, respectively.
14
Table 2.1 Summary of available patient data on mutants discussed in this thesis as well as the
average values for fALS and sALS cases. The higher numbers of patients for disease onset age
compared to duration are from diagnosed patients who are still alive.
Duration in
Onset in years
Mutant
years (number
(number of
Source
of patients)
patients)
a
A4S
>3 (1)
34 (1)
(63)
A4T
1.5 (21)
45.3 (26)
(64)
A4V
1.2 (205)
47.7 (212)
(64)
G37R
17 (27)
36.9 (27)
(64)
G41D
14.1 (15)
45.2 (17)
(64)
H48Rb
(32)
H80R
1.5 (1)
24 (1)
(65)
N86Db
(66)
I112T
0.9 (2)
44.0 (2)
(64)
I113T
4.3 (38)
56.3 (43)
(64)
V148G
2.1 (11)
43.1 (11)
(64)
V148I
fALSc
sALS
1.7 (5)
4.0
2.5 (269)
28 (4)
45
58.4 (359)
(64)
(67)
(64)
These mutations were introduced into the pHSOD1ASlacIq vector via polymerase chain reaction
(PCR)-site directed mutagenesis to enable the large scale expression of the mutants required for
in depth biophysical studies.
15
2.2
Methods
2.2.1 Production of recombinant fALS-associated hSOD plasmids using PCRmediated site-directed mutagenesis
Plasmids encoding fALS-associated hSOD mutants were expressed in the E. coli strain
XL1-Blue. A modification of the QuickChange Site Directed Mutagenesis protocol (Strategene,
La Jolla, CA) was used to produce the fALS-associated hSOD mutants A4S, A4T, G37R, G41D,
H48R, N86D, I112T, V148G and V148I and an sALS-associated mutant H80R. The primers
were designed by Joe Gaspar and Dr. Peter Stathopolous. Complementary primers for each
mutant were used to introduce the desired point mutation into the pHSOD1ASlacIq template
using PCR. Each primer was designed with the desired mutation about 10-15 bases from each
end of the primer and contains a minimum Guanine+Cytosine (GC) content of 40%. The primers
were also designed to end in 2-3 cytosines or guanines. The melting temperature (tm) for each
primer should be above 78 C. Table 2.2 lists the sequence modifications made for each mutant.
16
Table 2.2 Primer sequences for A4S, A4T, G37R, G41D, H48R, H80R, N86D, I112T and
V148G/I with %GC and tm. The first line in each primer shows the original sequence while the
second line shows the primers with the point mutations introduced (X - point mutation, X original nucleotide). tm of the primers were provided by Sigma Genosys Canada, Oakville,
Ontario.
tm
%
Mutant
Primers (5'-->3')
GC
(C)
1
GCG GCC ACA AAG GCC GTT GCT GTT TTG AAG GG
56.25
83.2
A4S
GCG GCC ACA AAG TCC GTT GCT GTT TTG AAG GG
GCG GCC ACA AAG GCC GTT GCT GTT TTG AAG GG
GCG GCC ACA AAG ACC GTT GCT GTT TTG AAG GG
56.25
83.2
G GTG TGG GGA AGC ATT AAA GGA CTG ACT GAA GGC CTG CAT GG
G GTG TGG GGA AGC ATT AAA CGT CTC ACT GAA GGC CTG CAT GG
54.76
86.4
C ATT AAA GGA CTG ACT GAA GGC CTG CAT GGA TTC CAT GTT C
C ATT AAA GGA CTG ACT GAA GAC CTG CAT GGA TTC CAT GTT C
43.90
79.8
C CTG CAT GGA TTC CAT GTT CAT GAG TTT GGA GAT AAT ACA GC
C CTG CAT GGA TTC CAT GTT CGC GAG TTT GGA GAT AAT ACA GC
47.62
83.1
CCA AAG GAT GAA GAG AGG CAT GTT GGA GAC TTG GGC AAT G
CCA AAG GAT GAA GAG AGA CGT GTT GGA GAC TTG GGC AAT G
50.00
83.0
G CAT GTT GGA GAC TTG GGC AAT GTG ACC GCG GAC AAA GAT G
G CAT GTT GGA GAC TTG GGA GAC GTG ACC GCG GAC AAA GAT G
56.10
87.6
CA GGA GAC CAT TCC ATC ATT GGC CGC ACA CTA G
CA GGA GAC CAT TCC ACC ATT GGC CGC ACA CTA G
57.58
82.4
V148G
CGT TTG GCT TGT GGT GTA ATT GGG ATC GCC C
CGT TTG GCT TGT GGT GGA ATT GGG ATC GCC C
58.06
84.1
CGT TTG GCT TGT GGT GTA ATT GGG ATC GCC C
CGT TTG GCT TGT GGT ATA ATT GGG ATC GCC C
51.61
78.9
A4T
G37R
G41D
H48R
H80R
N86D
I112T
V148I
1
2
Table 2.3 lists the components required for the PCR master mix. The buffer (Buffer 1) as well as
DNA polymerase mix was obtained from the Expand Long Template PCR System kit (Roche
Diagnostics, Indianapolis, IN). To decrease the effects of primer dimerization, the
complementary primers of each mutation were thermocycled separately in sterile PCR tubes,
containing 24.25 L of master mix and 0.25 L of one of the two complementary primers (final
17
concentration 0.25 mM) for five cycles before being pooled together. Thermocycling was done
using a TC-512 Thermal Cycler (TECHNE, Cambridge) using the conditions listed in Table 2.4.
The PCR products were then treated with DpnI endonuclease, which digests methylated or
hemimethylated DNA, for 1 hour in a 37 C water bath. Since the synthesized DNA is
unmethylated, DpnI will only digest the template plasmid. The recombinant plasmids were
stored at -20 C for further analysis.
Table 2.3 PCR components for master mix. Volumes listed are for final amount after
complementary tubes are pooled together.
Component
Volume
Final
added (L)
Concentration
Expand Long Template PCR System Buffer 1
5
pHSOD1ASlacIq template
0.5
~ 1.75 mM
dNTPs
0.5
0.25 mM each
MilliQ H2O
42.5
Total
48.5
Table 2.4 Conditions for thermocycling
PCR stages
Conditions
Initial denaturation
92 C for 5 min
Pause
Initial amplification
(5 cycles)
Pause
Final amplification
(15 cycles)
Final elongation
18
19
2.3
Results
2.3.1 Obtaining recombinant fALS-associated hSOD plasmids using PCRmediated site-directed mutagenesis
The PCR-mediated site-directed mutagenesis introduces desired mutations to the
pHSOD1ASlacIq template as well as amplifies the amount of recombinant mutant plasmid. The
PCR products were run on 0.7% agarose gels to determine if the process was successful. Figure
2.3 is a set of PCR products containing strong, single bands for each mutant, indicating that the
respective primers annealed with high specificity and that each sample was successfully
amplified. The mutant I112T was never successfully amplified (data not shown) while A4T was
obtained during my CHEM 494 with Dr. E. M. Meiering.
Figure 2.3 PCR products from PCR-mediated site directed mutagenesis of ALS-associated
hSOD mutants. The pHSOD1ASlacIq template is 5747 bp in size. Samples were run on a 0.7%
agarose gel at a constant voltage of 100 kV. *GeneRuler 1 kb DNA Ladder (Fermentas
Canada Inc., Burlington, Ontario)
20
Figure 2.4 SDS-PAGE of recombinant hSOD mutants obtained expression. Purified pWT
(lane A7 and B8) was used as a marker for hSOD.
2.3.3 Confirmation of recombinant hSOD mass
After it was determined that the recombinant mutants expressed hSOD, the mass of the
recombinant mutants were checked via mass spectrometry to verify if the correct mutation had
been introduced. Table 2.5 reports the final masses for the successfully obtained mutants and a
representative mass of the unsuccessful mutants. A protein of known mass was used as a
standard to obtain any offset the mass spectrometer might have. The mutants A4S, A4T, G37R,
G41D, H80R, V148G and V148I were successfully obtained. Due to the similarity in mass of
21
N86D to pWT, the N86D plasmid was sent for sequencing which revealed that the mutagenesis
was unsuccessful.
Table 2.5 Summary of recombinant hSOD
succefully obtained.
Apparent
Mutants
Offset (Da)
mass (Da)
A4S
15768.50
-1.99
A4T
15783.00
-1.42
G37R
15851.50
-1.99
G41D
15810.00
-1.99
H48R
15769.00
-1.42
H80R
15771.50
-1.99
N86D
15752.20
-1.99
a
15753.80
n/a
I112T
V148G
15711.00
-1.42
V148I
15767.00
-1.42
Expected mass
(Da)
15770.39
15784.42
15853.52
15812.42
15773.43
15773.43
15755.37
15742.34
15712.31
15768.41
Mutant
obtained (Y/N)
Y
Y
Y
Y
N
Y
N
N
Y
Y
A standard was not ran on the day I112T was checked, however, the apparent mass was sufficiently different from
the expected mass to conclude that the PCR-mediated site directed mutagenesis was not successful.
2.4
Discussion
PCR-mediated site directed mutagenesis is a useful method for introducing mutations to a
template plasmid. Of the 10 mutations tried, 7 were successfully obtained. The PCR attempts of
I112T and N86D resulted in pWT which could be due the 35 exonuclease proofreading
activity of the Tgo polymerase contained in the Expand Long Template PCR System DNA
polymerase mix. It is not clear why, despite several attempts, we have been unsuccessful in
obtaining I112T. On the other hand, the primers for H48R and N86D contain more than one
point mutation. It would be worth considering redesigning the primers for all 3 mutants to
increase the efficacy of mutagenesis. The characterization of the mutant proteins is described in
the following chapters. Chapter 3 will focus on determining the stability of holo A4S, A4T and
G41D and comparing them with pWT. This chapter will also include preliminary data on H80R.
22
In Chapter 4, the propensity for aggregation of holo A4S and A4T is studied to better understand
the mechanism by which these mutants may confer toxicity.
23
3.1
Introduction
PCDs are a fast growing subset of modern world diseases, largely due to the complicated
nature of the diseases and the lack of known causes. With many PCDs associated with natively
folded proteins, mutations associated with the hereditary forms of the diseases tend to destabilize
the native state (70). Similarly, studies on purified ALS-associated hSOD mutants have revealed
that the mutants are destabilised compared to pWT in the holo state (71-73). The destabilization
of the native state can in turn increase the proteins propensity to unfold/misfold or populate
aggregation prone intermediates. Mutant hSOD expressed in rabbit reticulocyte lysate have also
been shown to be less resistant to proteolysis (74) and hSOD peptide fragments have been found
in aggregates from a mutant hSOD mouse model of ALS (49). In this chapter, the thermal
stability of the dimer interface mutants A4S, A4T and I113T, in the holo form, will be
characterized using differential scanning calorimetry and compared to pWT. In addition, some
preliminary DSC data and analysis on G41D and H80R are presented.
3.1.1 Differential scanning calorimetry
The thermal stability of hSOD can be measured using differential scanning calorimetry
(DSC) (75). DSC measures the difference in heat capacity between a reference cell, containing
only buffer, and a sample cell, containing protein in buffer, as it is heated at a constant rate
through its unfolding transition (76). As the protein unfolds, more energy is required to heat the
sample cell at the same rate as the reference cell, which produces an endothermic peak in the
heat capacity curve (illustrated in Figure 2.1). The area under the unfolding transition peak gives
24
where
Equation 3.1
is temperature.
Figure 3.1 The specific heat capacity function of a globular protein. is the difference
between the heat capacity of the native ( , ) and the unfolded ( , ) states which are indicated
is the temperature where = 0.58 (figure
by dashed and dotted lines, respectively, and
obtained from (68)).
The vant Hoff enthalpy (
to
proposed model as well as the size of the cooperative unit of unfolding (76). In the case of
hSOD, unfolding can be fit to a dimer two-state unfolding model,
N2 2U
where the unfolding transition occurs between native dimer (N2) and unfolded monomers (2U)
(28).
For a dimer two-state unfolding model, the unfolding transition is expected to be
asymmetrical and the
unfolding model fits the data well. On the other hand, a ratio of greater than one suggests a
cooperative unit larger than dimer, e.g. the presence of aggregates, while a ratio of less than one
indicates a cooperative unit smaller than dimer, e.g. the presence of an intermediate (28,75,76).
The
correlates to the stability of the protein and most fALS-associated hSOD mutations have
been found to be less stable than pWT (28,68,77,78). The DSC trace also provides information
on the metallation state of the sample as mismetallated protein samples result in traces with
multiple peaks or shoulders at lower temperatures. In addition to having lower stability, most
fALS-associated hSOD mutants have a
indicates an increased propensity to aggregate in vitro (68,77,79). It is found that dimer interface
mutants A4S, A4T and I113T are destabilized compared to pWT in the holo state. They also
have slightly weakened dimer association compared to pWT, consistent with the location of
these mutations in the dimer interface. These findings agree with previous data on apo dimer
interface mutants, which also had weaker dimer association compared to apo pWT (77,79).
26
3.2
Methods
(25 mL/1 L culture) buffer containing sucrose (20% (w/v) final concentration) and
ethylenediaminetetraacetic acid (EDTA; final concentration of 15 mM), incubated in on ice for
20 min with gentle agitation and centrifuged at 4000 x g for 25 min. The cell pellet was then
resuspended in cold MilliQ (Millipore Ltd., Bedford, MA) water (20 mL/1 L culture) to release
the periplasmic proteins into solution, incubated on ice for 20 min with gentle agitation and
centrifuged at 4500 x g for 45 minutes. The supernatant was flash frozen in liquid nitrogen and
stored at -80 C until further purification.
3.2.2.2 Heat treatment and copper charging
The OS supernatant was diluted to a concentration of ~ 0.8 mg/mL in 20 mM Tris-HCl
pH 7 buffer and heated to 70 C with CuSO4 (final concentration of ~1.4 mM) for 25 mins. The
CuSO4 is added to ensure proper metallation of hSOD and the high temperature denatures most
other periplasmic proteins present in the osmotic shock mixture, which can then be removed via
centrifugation at 10 000 x g for 20 min. H80R was heated at 55 C due to its lower stability.
3.2.2.3 Hydrophobic interaction chromatography
The final purification step was carried out as described previously (68). Briefly, hSOD
was eluted from a hydrophobic interaction column using a high to low salt gradient. Fractions
containing hSOD were pooled together and dialyzed against MilliQ water with 4 exchanges over
24 hours using 6-8 kDa cutoff dialysis tubing (Spectra/Por molecularporous membrane tubing;
Spectrum Laboratories, Inc., Rancho Dominguez, CA). The dialyzed solution was concentrated
using an Amicon ultrafiltration device with a 10 kDa cutoff membrane (YM10 Amicon
ultrafiltration regenerated cellulose membrane; Millipore Ltd., Bedford, MA), filtered with a
0.22 m Acrodisc syringe filters (Pall Canada Ltd., QC) and stored at -80 C.
28
of superoxide and can be monitored spectrophotometrically (at 420 nm) by a colour change from
clear to yellow-brown. The addition of hSOD to pyrogallol will inhibit pyrogallols
autooxidative ability since hSOD competes for superoxide. The specific activity of hSOD will be
measured in Units/mg where 1 unit is the amount of hSOD required to reduce the rate of
pyrogallol autooxidation by 50%. The average specific activity of holo pWT is approximately
1800 U/mg (73). Many mutants retain full activity; however, a lower than expected specific
activity could indicate that the protein is not properly metallated. It may also indicate that a
fraction of the protein in the sample is not functional, or the sample as a whole has lower
dismutase activity, or a combination all three causes.
3.2.5 Differential scanning calorimetry
Measurements were made using methods previously established in the Meiering
laboratory (28,68). Briefly, samples were prepared in 20 mM HEPES pH 7.8 and degassed prior
to loading into the calorimeter. All measurements were made using a MicroCal LLC VP-DSC
(MicroCal Inc., Northampton, MA) and scanned versus dialysate. Initial DSC scans of holo and
apo proteins were done from 15-100 C and 10-90 C, respectively, at a rate of 1 C/min.
The data were then fit to a dimer 2-state unfolding model (Equation 3.2) as well as a
monomer 2-state unfolding model (Equation 3.3) as described elsewhere (75,77,79) using the
following equations:
1
Equation 3.2
Equation 3.3
30
where
and
is a
and
is the
unfolded baselines, respectively. All DSC data fitting was done using Microcal Origin 5.0. The
fit provides a fitted
= 0.5), a
value and a
that is the
multiplying
determine the
. The DSC scans for holo hSOD were fit to 75% of the endothermic
peak due to the presence of an exotherm at high temperatures. The exotherm is probably from
irreversible protein aggregation at the high temperatures so that the unfolded baseline is not well
defined (77).
For a 2-state unfolding process,
1 ln
constant
Equation 3.4
molecularity of the unfolding subunit . Rearranging equation 3.4 indicates that the slope,
ln versus 1
is the
, for a
plot is
Equation 3.5
which can then be used to calculate the molecularity of the unfolding subunit.
31
3.3
Results
deviates from the thermograms, where the fits predict a more symmetrical endotherm compared
s and at the beginning of the endothermic peak.
The concentration dependence of the dimer interface mutants is illustrated in Figure 3.3
where the
gradually increases with increasing protein concentration. Figure 3.4 are plots of
calculated from the average fitted parameters to the dimer 2-state unfolding model from Table
3.1. Figure 3.4 shows that the concentration dependence of the experimental
A4S, A4T and I113T fall within the predicted concentration dependence of their
protein concentration range. All three mutants have experimental
predicted
s of holo pWT,
s at the lower
s at higher protein concentrations. This may be due to their lower stability, leading
parameters for holo pWT, A4S, A4T and I113T. Holo pWTs
holo pWT fits the dimer 2-state unfolding model well. The ratios for A4S, A4T, A4V and I113T
are 0.850.07, 0.950.13, 0.720.18 and 0.890.07, respectively, which are quite close to 1. The
slightly decreased values could be indicative of the formation of a monomer intermediate, which
would be expected since the mutations are near the dimer interface. The
mutants also increase with protein concentration, with
higher protein concentrations. This indicates that the dimer is more populated at higher protein
concentrations.
From Equation 3.5, the molecularity of the unfolding subunit can be calculated using the
following equation:
Molecularity
where
is the average
lnP versus 1/
Equation 3.6
plot. Table 3.2 list the molecularity for holo pWT and dimer interface mutants
as well as dissociation constants, Kd, of the dimer interface mutants in the holo and apo state. The
molecularity for pWT is 1.91, which is close to the expected value of 2 for dimer unfolding. On
the other hand, the dimer interface mutants all have lower molecularities compared to pWT,
consistent with their low
in Table 3.1.
33
Figure 3.2 Dimer 2-state and monomer 2-state fits of holo dimer interface mutants A4S (A
and B), A4T (C and D), I113T (E and F) and G41D (G and F) at protein concentrations of 0.5
mg/mL, 0.39 mg/mL, 0.7 mg/mL and 1.00 mg/mL, respectively. The dimer 2-state fits are on the
left while the monomer 2-state fits are on the right. The experimental data and the fits are
represented by solid and dashed lines, respectively.
34
Figure 3.3 Protein concentration dependence of holo pWT and holo dimer interface
mutants. DSC thermograms of pWT (A), I113T (B), A4S (C) and A4T (D) at different
concentrations are offset for clarity and arranged in increasing protein concentration from bottom
to top. The experimental data and the dimer 2-state fits are represented by solid and dashed lines,
respectively. The pWT data were obtained by Dr. Jessica Rumfeldt (68).
35
36
Table 3.1 Fitted dimer 2-state parameters for holo dimer interface mutants A4S, A4T, A4V
and I113T and holo pWT.
SOD
[SOD]
(mg/mL)
pWT*
0.20
pWT*
0.50
pWT*
0.50
pWT*
1.00
pWT*
1.87
pWT*
6.30
MeanS.D.
A4S
0.10
A4S
0.30
A4S
0.30
A4S
0.45
A4S
0.50
0.95
A4S
A4S
1.00
2.95
A4S
MeanS.D.
A4T
0.16
A4T
0.20
A4T
0.38
A4T
0.39
A4T
0.44
A4T
0.59
A4T
1.08
A4T
4.34
MeanS.D.
A4V
0.20
0.50
A4V
1.00
A4V
5.00
A4V
MeanS.D.
I113T
0.09
I113T
0.70
1.48
I113T
I113T
2.76
MeanS.D.
G41D
1.00
*
b
c
d
(C)
91.60.2
92.40.4
91.60.5
92.50.5
93.20.3
94.20.3
(kcal/mol)
265.515.1
256.522.8
230.624.3
269.030.5
276.920.5
238.720.2
(kcal/mol)
228.811.5
231.418.3
220.620.9
255.026.0
270.517.8
279.820.4
86.70.0
86.50.2
87.80.1
87.30.0
88.20.0
88.50.0
87.80.0
89.10.0
208.23.6
184.410.1
234.27.1
212.93.8
238.33.0
238.25.2
211.13.1
214.92.5
256.13.1
243.111.2
276.87.0
217.02.1
279.82.1
261.83.1
253.51.7
260.11.3
85.20.1
85.40.1
85.60.0
85.80.0
86.20.1
86.60.0
86.20.1
87.30.0
187.94.0
197.36.3
209.82.2
201.52.1
215.47.7
230.32.0
221.54.6
224.11.6
254.84.2
195.55.0
185.21.0
251.41.3
243.67.6
239.50.9
207.33.4
229.90.8
86.90.0
87.20.0
87.40.0
88.40.0
197.71.4
204.21.4
208.41.7
237.53.1
373.91.5
333.81.0
232.00.8
279.91.5
86.80.0
88.10.0
88.20.0
89.00.0
200.72.0
242.02.3
231.62.1
260.01.6
253.81.4
248.21.6
257.11.5
288.30.8
85.30.4
235.85.7
225.2.74.4
1.16
1.11
1.05
1.05
1.02
0.85
1.040.10
0.81
0.76
0.85
0.98
0.85
0.91
0.83
0.83
0.850.07
0.74
1.01
1.13
0.80
0.88
0.96
1.07
0.97
0.950.13
0.53
0.61
0.90
0.85
0.720.18
0.79
0.97
0.90
0.90
0.890.07
1.05
c
at
(kcal/C/mol)
4.12
4.50
5.37
5.80
7.07
3.73
5.101.23
3.85
2.92
3.57
3.95
5.02
11.21d
4.32
4.60
4.030.69
1.74
0.04
3.78
3.33
8.13
6.22
4.77
4.34
4.042.51
5.88
6.49
3.44
4.88
5.171.33
3.42
4.19
5.51
5.53
4.661.04
6.13
37
Table 3.2 Dissociation constants, Kd, and molecularity for pWT and dimer interface
mutants. The molecularity was calculated using equation 3.4.
hSOD
Kd apo
37 C a
( M)
Kd holo
65 C b
( M)
Molecularity
pWT
1.91
A4S
1.60.6
1.86
A4T
5.11.2
1.57
A4V
10.03.1
4817
1.39
I113T
3.90.1
3310
1.54
a
b
to pWT range from -4.7 C to -7.4 C. The negative values of (Table 3.3), the change in
of the mutant relative to pWT, and the shift to lower temperatures of the free energy diagrams of
the mutants relative to pWT (Figure 3.3 B) also indicate that the mutants are destabilized
compared to pWT. Due to errors that may be propagated through long extrapolations of , the
is reported at
concentrations are more representative of dimer 2-state unfolding and will provide a more
relevant comparison with pWT values. A constant
previously for pWT by Dr. Jessica Rumfeldt (68)) instead of the average
the
s from the DSC fits are associated with considerable error (85) owing to the lack of a
equations (68,77):
Equation 3.7
Equation 3.8
Equation 3.9
Equation 3.10
39
Figure 3.5 ALS-associated mutants are destabilized relative to pWT in the holo state. (A)
DSC thermograms of holo pWT, I113T, A4S, A4V, A4T and G41D at protein concentrations of
1.00 mg/mL, 0.70 mg/mL, 1.00 mg/mL, 1.00 mg/mL, 1.08 mg/mL and 1.00 mg/mL,
respectively. The thermograms are arranged from highest to lowest
(top to bottom) where the
solid lines are experimental data and the dashed lines are the fits to the dimer 2-state unfolding
model. The thermograms are also offset for clarity. (B) Free energy diagrams of each mutant
relative to pWT in the holo state. pWT is shown as a dashed line while A4T, A4S, A4V, G41D
and I113T are shown in solid lines and indicated with arrows. The free energy plots were
calculated from average fitted values from higher protein concentration datasets listed in Table
3.1 as they are more representative of dimer 2-state unfolding. The intersection of the horizontal
line, -RTlnP (for 0.50 mg/ml), with the free energy plots gives predicted tm for each mutant.
40
Table 3.3 Thermodynamic parameters of holo fALS-associated mutants A4S, A4T, I113T and G41D compared to pWT. Only
datasets at higher protein concentrations (Table 3.1 ) for the dimer interface mutants were used in the calculation for the parameters
in this table to avoid any bias that may be contributed from the presence of a monomer intermediate.
b
Holo
(25 C)
(25 C)
(25 C)
a
c
d
isomer
(C)
(kcal/mol) (kcal/K/mol) (kcal/mol) (kcal/K/mol)
(kcal/mol)
(kcal/mol) (kcal/mol) (kcal/mol)
*
pWT
10.10.4
0.660.05
35.43.2
0.130.05
A4S
-4.8
6.80.3
0.590.04
28.12.8
0.060.04
-3.3
-7.4
-0.06
-26.7
A4T
-6.5
5.70.5
0.620.01
28.71.2
0.090.01
-4.4
-6.8
-0.04
-17.9
A4V
-5.8
6.20.4
0.600.05
28.12.9
0.070.05
-3.9
-7.4
-0.05
-23.7
I113T
-4.7
6.80.2
0.660.04
32.12.3
0.130.04
-3.3
-3.3
0.00
-3.5
G41D
-7.4
4.9
0.66
30.60.5
0.130.00
-5.1
-4.9
0.00
-3.6
a
).
c
Change in the of the mutants relative to pWT (
).
d
Change in the of the mutants relative to pWT (
).
*
Data from (68).
41
41
3.3.2 H80R
3.3.2.1 Initial Data
When H80R was initially obtained from the small scale purification, the mass spectrum
of its OS contained not only a peak corresponding to H80R but also a fairly large peak
corresponding to an approximately +18.5 Da species. Another mass spectrum was obtained after
a 16 days incubation period at room temperature. Figure 3.6 illustrates the peak intensities of
H80R and the +18.5 Da peak before and after the 16 day incubation. Over time the +18.5 Da
species increased as the H80R peak decreased. This indicated that H80R was being modified
over time. Figure 3.6 also illustrates that this modification process began either before or during
the protein purification process as the pre-incubation sample contained the modification. The
addition of 18.5 Da could correspond to the mass of a molecule of water, suggesting that the
42
protein may have been cleaved somewhere between the intramolecular disulfide bonds (C57C146) of H80R, thus keeping the cleaved protein intact.
42
Figure 3.6 Change in H80R species distribution over time. The spectrum indicates conversion
of H80R into H80R+18.5 kDa species. The sample is the OS supernatant form a small scale
purification of H80R.
decreased stability of H80R due to the lack of zinc coordination (86), increasing its sensitivity to
proteolysis.
To determine where the protein was being cleaved, the OS from a small scale expression
was purified using a Nanosep 3K Omega centrifugal device (Pall Canada Ltd., QC) to remove
the sucrose as well as any small fragments that may be in the sample. The OS and the slightly
purified sample were incubated for 17 hours at 37 C and at room temperature. Figure 3.7 is an
SDS-PAGE of H80R incubated at the different temperatures. The unreduced samples (Figure 3.7
Lanes 9-14) contain 2(3) major bands of ~15 kDa in size. We suspect that the larger band(s)
corresponds to the cleaved species as it would have a more expanded structure and travel slower
through the gel. The same samples reduced with 10% -mercapthoethanol (me) (Figure 3.7
Lanes 3-7) reveal that one of the larger bands has been replaced by 2 smaller fragments of ~8.1
kDa in size.
The specific masses of the fragments were determined using mass spectrometry. sample
was fully reduced. Figure 3.8 shows the reduced (B) and non-reduced (A) mass spectra of the
slightly purified H80R that had been incubated for 17 h at 37 C. To ensure that the entire
sample in Figure 3.8 (B) was reduced, the sample was incubated overnight under reducing
conditions in 1% me. Figure 3.8 (A) is the spectrum for the non-reduced sample and contains
peaks for disulfide intact H80R (15772 Da) and disulfide intact H80R+18.5 Da (15790.5 Da)
while Figure 3.8 (B) is the spectrum for the reduced sample and contains reduced H80R
(15774.5 Da), no H80R+18.5 Da peak and peaks at 7211 Da and 8580 Da. Using the MassLynx
program, it was determined that the 7211 Da fragment corresponds to residues 1-69 while the
8580 Da fragment corresponds to residues 70-153. Therefore, H80R was being cleaved between
44
residue R69 and K70. Both spectra in Figure 3.8 contain a ~H80R+98 Da species. This +98 Da
peak is present in most of our protein preparations as is thought to correspond to a sulphate
adduct (artefact from our protein expression and purification protocol).
45
Figure 3.8 Mass spectrum of fragmented H80R in non-reducing (A) and reducing (B)
conditions. The H80R sample was incubated at 37 C for 17 hours before the mass spectrum
was obtained, at which point, an aliquot of the sample was reduced with 1% me. Both aliquots
were sprayed in a solution of 1:1 acetonitrile:water mixture with 0.2% formic acid.
PAGE illustrating the results for the different H80R expression and heat treatment conditions.
H80R grown with zinc had the highest yield compared to the cultures grown with both copper
and zinc or without metals, suggesting that the addition of zinc helps stabilize H80R. Figure 3.9
also indicates that heat treatment without the addition of copper does not significantly increase
the purity of the OS. Samples heat treated with copper at 70 C (Figure 3.9 Lanes 6, 13 and 19)
have a decreased amount of H80R compared to the samples heat treated at lower temperatures,
indicating that H80R likely aggregates at 70 C in the presence of copper. Interestingly, reduced
samples from all three sets of OS that were heat treated with copper (Figure 3.9 Lanes 21-23)
contain a smaller fragment that does not show up in their unreduced counterparts (Figure 3.9
Lanes 4, 11 and 17), further implicating coppers role in the cleavage of H80R. In light of these
results, H80R preparations were expressed with zinc, for higher yield, and , initially, not heat
treated prior to purification on the hydrophobic interaction column as heat treatment without
copper did not increase the purity of the OS. The apo H80R was made from purified as isolated
preparations that were heat treated with copper at 55 C. This was due to the low yields from the
non-heated treated samples attributed to protease activity during the protein purification process
as well as the non-ideal elution profile of non-heat treated H80R on the hydrophobic interaction
column. All the as isolated H80R preparations were checked using SDS-PAGE with me after
purification and contained only whole H80R.
47
Figure 3.9 SDS-PAGE of H80R grown and heat treated with different conditions to
determine the most optimal set of expression parameters. H80R was grown without metals (1-6),
with both copper and zinc (8-13) and with only zinc (14-19). The osmotic shock solutions from
these were then heat treated with or without copper at 50 C, 60 C and 70 C. Samples in lanes
20-23 were run with 10% me. Lanes 21-23 appear to have a low molecular weight fragment
that is not present in lane 20, or any other lanes.
48
of ~59 C. However, all the thermograms, with the exception of (d), contain a shoulder with a
of ~74 C. When the samples were rechecked using SDS-PAGE with me, the samples
contained a small amount of fragments similar to those observed in Figure 3.7. This suggested
that the fragmentation process was either occurring while the samples were stored at -80 C or
that it was a fast process occurring within the time required to prepare the samples. The shoulder
at 74 C is perhaps due to the unfolding of aggregates of the fragmented H80R. Despite the
presence the fragmented species of H80R, the
concentration dependence of apo H80R over a concentration range 0.76-1.43 mg/mL suggests
that apo H80R may be unfolding via a monomer 2-state unfolding model.
49
Figure 3.10 DSC thermograms of as isolated (d and e) and apo (a-c) H80R. The three apo
samples, a, b, and c, were at protein concentrations of 1.43 mg/mL, 0.97 mg/mL and 0.76
mg/mL, respectively, while the as isolated samples, d and e, were at protein concentrations of
0.46 mg/mL and 0.43 mg/mL.
50
3.4
Discussion
3.4.1 Dimer interface mutants and G41D are destabilized compared to pWT in the
holo state
The thermal stabilities of the mutants relative to pWT in the holo state were determined
using DSC. The mutants have comparable specific activities, with the exception of G41D, to
pWT which has a specific activity of 1800200 U/mg (77). The dimer interface mutants A4S,
A4T and I113T have average specific activities of 194363 U/mg, 1736260 U/mg and 1668
U/mg, respectively, while G41D has a specific activity of 1538 U/mg ( S.D. of different batches
of protein). However, the thermogram for G41D does not appear to contain multiple peaks or
shoulders, suggesting that the sample is homogeneous. Alanine 4 is located in -strand I, at the
edge of the dimer interface, and has its side chain pointing into the barrel. On the other hand,
isoleucine 113 is located in loop 6 and points into the dimer interface. The high specific activities
of the dimer interface mutants are not unusual as they are located, structurally, quite far away
from the metal binding region. It has also been found that most non-metal binding region
mutations do not affect the activity of hSOD (19). The lower activity of G41D may be due to its
close proximity to the MBR. Glycine 41 is located at the beginning of -strand IV, which then
extends to the zinc binding loop. The introduction of a charged residue at the edge of the barrel
could also perturb the integrity of the barrel, which in turn would affect the structure of the
active site. Interestingly, recent computational studies indicated that mutations located away
from the dimer interface or MBR affect the integrity of both the dimer interface and MBR(36).
All the mutants are destabilised relative to pWT in the holo form as reflected by their
lower
s and s. This is consistent with previous data obtained on holo hSOD mutants where
51
mutants were shown to be destabilized via thermal unfolding (28,71)((72) Chapter 4) ((73)
Chapter 2) as well as chemical denaturation (78)((68) Chapter 2). The lower
dimer interface mutants are due to lower
Previously studied mutants (A4V, G85R, G93A/D/R/S/V and E100G) are also enthalpically
destabilized in the holo state (68,77). Interestingly, G41D and I113T have values of 0 and
only a small decreases in
not a dimer interface mutant. It is located at the beginning of -strand IV and also replaces a nonpolar residue with a charged residue. In fact, G41D is one of the few mutations in hSOD that
increases the net negative charge of the protein as the majority of ALS-associated mutations
decrease the net charge of hSOD (87). It is worth noting that the dimer interface mutants are
associated with short disease durations while G41D has relatively long duration (Table 2.1)
and has comparable
at physiological
temperature relative to the dimer interface mutants. The long duration of G41D could be
attributed to the increase in net charge of the protein which is predicted to decrease aggregation
by increasing electrostatic repulsion the protein molecules (88).
3.4.2 Dimer interface mutants have decreased dimer stability
Although the dimer interface mutants are well fit to the dimer 2-state model, their low
/
action, the monomer intermediate would be more populated at low protein concentrations while
the dimer would be more populated at high protein concentrations (79). This change in
/
protein concentration and can be observed for the dimer interface mutants (Table 3.1), where the
52
lower
ratios tend to correlate with lower protein concentrations. This trend is found
in all the holo dimer interface mutations studied in our laboratory which includes A4S, A4T,
A4V (68) and I113T as well as non-dimer interface mutants including E100G (68), G93A and
G93R (77). However, monomer 2-state fits for the dimer interface mutants and G41D show
systematic deviations from the DSC thermogram similar to those observed for pWT (28),
indicating that these mutants do not unfold in a purely monomer 2- state manner. It in worth
noting that apo A4V and H46R have been shown to thermally unfold in a 3-state monomer
intermediate unfolding mechanism (79). In addition, chemical denaturation studies on holo pWT,
G85R, E100G and G93A and apo G85R, G93R, E100G and I113T reveal that they undergo a 3state monomer intermediate guanidinium chloride-induced denaturation (68,72,78).
Additionally, isothermal titration calorimetry (ITC) on holo A4V and I113T at 65 C
indicated that these mutants had weaker dimer association compared pWT. ITC of holo A4V and
I113T gave dissociation constants (
not give significant heats of dissociation (quantifiable dissociation heats for the only first ~2
injections), and therefore, a
did not give significant heats of dissociation. However, the G93 mutants had more injections
with quantifiable heats of dissociation compared to pWT, indicating that they have a weaker
dimer interface compared to pWT (77). ITC experiments conducted on apo hSOD at 37 C also
provide evidence for weaker dimer association in ALS-associated mutants (79). Similar to holo
pWT, apo pWT did not give significant heats of dissociation at 37 C while dimer interface
mutants A4S, A4T, A4V and I113T gave
respectively (79). On the other hand, non-dimer interface mutants H46R and G93S gave
53
s of
0.5 M and 1.2 M, respectively (79). It is not surprising that apo A4V had the highest
out of
all the mutants as DSC indicated that A4V unfolds via a monomer intermediate. This is because
the dimer interface is sufficiently destabilized compared to monomer stability, enabling the
detection of the two processes in DSC. On the other hand, apo H46R which also unfolds with a
monomer intermediate had the highest measurable
monomer stability coupled with a dimer interface that is only mildly destabilized compared to
pWT (79), which also enables the detection of the two processes in DSC. Accordingly, the
monomer intermediate is never significantly populated if a mutant has a strong dimer interface or
a destabilized monomer. In light these results, it would be sensible to investigate if holo A4S,
A4T and I113T undergo a 3-state monomer intermediate unfolding mechanism.
3.4.3 H80R
The presence of fragments in all the preparations of H80R regardless of the purification
conditions is perplexing and the reason for this has yet to be resolved. It is possible that aberrant
copper chemistry is responsible for the fragmentation of H80R. The formation of fragments after
storage in -80 C suggests a metal catalyzed process. Mutant hSODs have been shown to
participate in non-native chemistry in vitro (40,53) and although copper was not added to the
initial as isolated preparations of H80R, the samples may contain small amounts of copper from
the growth media. Figure 3.11 illustrates the close proximity of residues R69 and K70, the
cleavage site, to the active site of hSOD as well as H80R, the mutation site. On the other hand, a
study conducted by Bruns et al. indicated that H80R has decreased resistance to proteolysis (74).
The cleave site is located in the zinc binding loop which, in the absence of zinc coordination, has
been shown to be disordered (86). An unstructured zinc loop in the mutant may have an
enhanced susceptibility to proteolysis.
54
Figure 3.11 Site of cleavage in H80R. Prepared using MolMol (29) and PDB coordinates 1SOS
(30). Each monomer contains a Zn2+ and a Cu2+ ion depicted as black and orange spheres,
respectively. Inset is a magnification of the metal binding sites as well as the R69-K70 site where
cleavage occurs.
3.4.3 Correlations with disease
Since the discovery of a genetic link between hSOD and ALS (21), many studies have
been focussed on finding the common denominator of hSOD toxicity. Due to the high stability of
the holo hSOD, including the mutants (89), many studies have suggested that the toxic form of
hSOD is the more destabilized apo form (90) or the disulfide reduced apo form which is
proposed to be the most immature form of hSOD in the cell (91). However, it has been found
that destabilization of the apo protein is not a common feature for all fALS-associated mutants
(92). A recent article by Wang et al. revealed a correlation between disease durations and the
sum of aggregation propensity and instability of the mutants (64). The correlation, although
55
compelling, only analyzes 28 different mutants due to the limited amount of data available.
Considering the complicated nature of ALS, it is likely that there is no common denominator and
that the different mutants associated with ALS confer toxicity via different mechanisms, all of
which may lead to increased population of some sort of destabilized species that has increased
propensity to aggregate compared to holo WT.
56
Introduction
As in many other neurodegenerative diseases, protein aggregation is a hallmark of ALS
pathogenesis (2). In vitro studies have shown that fALS-associated hSOD mutants have an
increased propensity to aggregate (73,93) while x-ray diffraction and nuclear magnetic resonance
studies have shown that the partially metallated, as well as fully metallated, S134N hSOD mutant
tend to generate protein oligomers (94,95). More importantly, aggregates from neuronal cell line
(13) and mice model studies of fALS (39,96,97) as well as fALS and some sALS patients have
been found to contain hSOD (14,50,51). It is not known if the hSOD found in these aggregates
retains their metal ions; however, given that the major species of hSOD in the cell is in the holo
form (77), it is plausible that aggregation could occur from holo hSOD.
Various studies on hSOD aggregation have focused on either the apo (90) or
mismetallated forms of hSOD, or on the reduced form (91), considering holo hSOD too stable to
give rise to aggregates in ALS (89). A recent study by Banci et al. demonstrated that disulfide
intact apo wildtype hSOD formed aggregates at close to physiological conditions of pH 7, 37 C
and a protein concentration of 100 M, while both holo wildtype hSOD and holo pWT did not
(90). The data from this study contradicts the results previously obtained by the Meiering lab
where holo pWT as well as holo forms of fALS-associated mutants have been shown to form
aggregates when incubated without agitation at physiologically relevant conditions of pH 7.8 and
37 C (77) (Y. M. Hwang unpublished data). This chapter investigates the aggregation properties
of holo dimer interface mutants A4S and A4T and possible correlations to thermal stability and
57
disease duration. The aggregation profiles indicate that the mutants have increased aggregation
propensities and also provide information regarding the possible aggregation mechanisms. In this
chapter, dynamic light scattering is used to monitor aggregation of dimer interface mutants A4S
and A4T as well as pWT in the holo state.
4.1.2 Dynamic light scattering
Dynamic light scattering (DLS) is one of the tools commonly used to monitor protein
aggregation (98,99). DLS utilizes the principles of Brownian motion to measure the size of
particles in a sample solution. A detector located at a fixed angle and distance with respect to an
incident light beam, picks up light scattered by particles in the sample solution, which is located
between the detector and the light source. The intensity of scattered light fluctuates about an
average intensity and the time dependence of these fluctuations are related to the diffusion
coefficient,
diffusion coefficient reflects the speed of particle movement and is inversely related to particle
size as described in the Stokes-Einstein equation:
Equation 4.1
where
The light scattering measurement is then divided into small time intervals or delay times,
, which are much shorter than the time required for the fluctuation to return back to the average
scattered intensity. The autocorrelation function,
as a function of time between . With increasing delay times, the correlation between the
58
is the
where
Equation 4.3
beam and
is the scattering angle. Because larger particles in a sample move slower through the
sample solution, they have smaller diffusion coefficients, which results in a slower decay. Figure
4.1 shows the autocorrelation function for a larger and a smaller particle.
59
Dynamic light scattering can be used to monitor particle size changes over time and has
been successfully used to monitor protein aggregation in vitro (77,100). In the case of samples
with more than one particle size distribution, the autocorrelation function will contain
exponentials from each particle size distribution. Algorithms such as cumulants (101) and
CONTIN (102) deconvolute the different exponentials and provide the hydrodynamic diameter
of the particles in solution.
4.1.2 Nucleation-dependent protein aggregation
Many disease-associated proteins have been shown to aggregate through a nucleationdependent process involving three phases: lag, growth and plateau (10,103). This process is
associated with a sigmoidal aggregation profile (Figure 4.2) and a lag phase that is associated
with the formation of thermodynamically unfavourable nucleus (104,105). Nucleation-dependent
protein aggregation typically is also characterized by a protein concentration dependence of the
lag phase whereby the lag phase shortens with an increase in protein concentration (104).
Another feature of nucleation-dependant protein aggregation is a seeding-effect, whereby the
addition of preformed aggregates eliminates the lag phase or nucleation step required for
aggregate formation (10,103). Both the concentration dependence and seeding effect have been
shown to occur in in vitro holo hSOD aggregation, indicating that holo hSOD aggregates via a
nucleation-dependent process ((77) and Y. M. Hwang unpublished data). However, the specific
pathways and mechanisms by which aggregation occurs are still being elucidated. In this chapter,
aggregation of dimer interface mutants in the holo form are monitored using DLS.
60
Figure 4.2 A graphical representation of the sigmoidal increase in mean light scattering
intensity upon aggregate formation (reproduced from (105)).
61
4.2
Methods
precision cell, type 105.251-QS, Hellma GmbH and Co., Mullheim, Germany). Samples were
gently pipetted prior to transfer into the cell and measurements were made twice a day during the
lag phase, 3 times a day during the exponential phase and once a day during the plateau phase.
62
The aggregation profiles obtained from both instruments agree well and, to avoid redundancy,
only aggregation time courses from the 90Plus Particle Sizer will be presented.
4.2.4 Determining lag time from fits to a sigmoidal function
The kinetics of holo hSOD aggregation were fit using Microcal Origin 5.0 to an empirical
sigmoidal function given by the following equation (also shown in Figure 4.2) (105):
Equation 4.5
where
is time and
is the time
constant which is the reciprocal of the apparent rate constant for the growth phase of the curve
(1/
). Although specific kinetic schemes cannot be obtained from the sigmoidal fit, it will
provides a way to quantitatively compare the aggregation profiles of the different hSOD mutants.
The lag time (duration of the lag phase) of the aggregation curve can be calculated from the
fitted parameters using the following equation:
Lag time
Equation 4.6
associated hSOD mutations undergo. Based on protein aggregation kinetics described by Ferrone
(104), the aggregation profile for primary-nucleation will fit to a cosine function:
fraction of completion
where
cos
Equation 4.7
the aggregation profile for secondary nucleation will fit a cosh function:
fraction of completion
cosh B
Equation 4.8
These fits are only applicable to the initial phase of the aggregation process, therefore, the
aggregation profiles of hSOD were converted to fraction of completion and only the first 20% of
the profiles were fit.
4.3
Acknowledgements
I would like to thank Young-Mi Hwang who assisted in collecting holo aggregation data
and Helen Stubbs who provided some holo pWT protein as well as assisted in collecting some
data.
4.4
Results
late seeds collected at the plateau phase. The lag phase of holo pWT is reduced from 200 h to 28
h by the addition of early seeds and completely eliminated by the addition of late seeds,
demonstrating that holo hSOD aggregation is nucleation dependent. These results agree with and
confirm previous results obtained for other holo hSOD variants (77) (Y.M. Hwang unpublished
data).
that did not match experimental initial mean light scattering intensities were considered preseeded and were not included in the lag time determination. Figure 4.4 illustrates the
aggregation profiles for pre-seeded and a non pre-seeded (successful) holo pWT samples. The
experimental initial mean light scattering intensities of both profiles are approximately 2.6 kcps,
however, the fitted initial light scattering intensities of the pre-seeded and non pre-seeded time
courses are -14.10 kcps and 2.66 kcps, respectively. The lower than expected fitted initial light
scattering intensity as well as the absence of an appreciable lag phase indicates that the sample
contains some preformed aggregates, perhaps as an artefact from filtering.
Figure 4.4 Pre-seeded and non pre-seeded 10 mg/mL holo pWT aggregation profiles. The
pre-seeded data points are in open circles while the non pre-seeded data points are in open
squares. The dashed lines are fits to Equation 4.5. The initial mean light scattering intensities, ,
of the pre-seeded and non pre-seeded sample are -14.10 kcps and 2.66 kcps, respectively.
Protein aggregation studies have shown that the combination of adsorption to
hydrophobic interfaces (filter membrane or air-water interface) and the shear stress associated
66
with the filtering process could exacerbate protein unfolding (106,107), which can lead to the
formation of aggregation nucleus. However, light scattering data from multiple holo hSOD
samples, prepared is a consistent manner, indicate that most of the samples are not pre-seeded
and the sample solution at the beginning of the time course as well as during the lag phase is in
fact monodispersed. Figure 4.5 is a representative intensity plot of a holo aggregation sample
over time. At t=0 h throughout the lag phase, the solution is monodispersed with a particle size
of ~5 nm, consistent with the size of hSOD dimer (77). As aggregation occurs, a larger species of
~1000nm appears.
Figure 4.5 Size distribution of holo hSOD aggregation over time. (A) Aggregation profile of
holo A4T with experimental data in solid squares and sigmoidal fit in dashed lines. (B) Intensity
plots at different time points corresponding to points indicated in (A) for holo A4T aggregation.
Aggregation of holo hSOD samples are quite variable with varying lengths of lag times,
growth phases and final mean light scattering intensities. This variability can be observed in
Figure 4.6. Nevertheless, Figure 4.6 (D) and Table 4.1 show that the dimer interface mutants
A4S and A4T appear to have shorter average lag times compared to pWT. The average lag time
of A4T, 77.7019.04 h, is clearly shorter than pWT, 135.2828.48 h, while the averaged lag
67
time of A4S, 109.4182.53 h, is slightly shorter than pWT. In Figure 4.6 (B), the aggregation
profiles in blue and red (indicated by arrows) are from a single holo A4S sample that was split
into two samples at the beginning of the experiment. However, they do not have similar lag times
or final mean light scattering intensities, indicating that the variation between samples are not
due to differences in sample quality but is intrinsic to hSOD aggregation.
Figure 4.6 Aggregation profiles of holo dimer interface mutants and pWT. Experimental
data for pWT (A), A4S (B) and A4T(C) are in hollow circles while the fits to Equation 4.5 are in
dotted lines. (D) Bar graph of the average and standard deviation of fitted lag times reflect the
variability in holo hSOD aggregation. A sample of holo A4S was divided into two identical
fractions and monitored in parallel. The arrows indicate the aggregation profiles for the two holo
A4S samples, which are in red and blue.
68
Table 4.1 Summary of holo hSOD aggregation fitted parameters to Equation 4.5.
Holo
Initial mean
Final mean light
lag time (hrs) b
a
a
scattering
hSOD light scattering
(hrs)
(hrs)
intensity (kcps) a intensity (kcps) a
pWT
pWT
pWT
pWT
pWT
2.190.48
2.680.29
2.660.09
2.650.30
-14.1012.24
17.821.77
15.730.86
4.830.08
7.330.30
29.111.77
222.1912.94
49.699.42
237.906.74
34.785.26
115.112.43
6.092.18
155.293.20
4.134.22
57.3066.90
130.7134.11
MeanS.D.
122.8216.01
168.348.55
102.933.27
147.045.30
-204.1275.10
13528
A4S
A4S
A4S
A4S
2.330.10
2.370.23
2.260.15
2.030.27
3.830.13
7.840.20
4.570.10
10.000.53
107.440.00
0.260.00
60.071.26
4.091.27
70.393.25
9.252.76
405.0921.58
89.0820.25
MeanS.D.
106.930.00
51.891.79
51.884.26
226.9329.59
10982
A4T
A4T
A4T
2.790.23
2.570.33
-33.7877.68
8.850.27
11.520.39
18.281.93
74.292.04
5.031.55
105.338.12
7.084.62
-101.11345.25
140.3176.10
MeanS.D.
64.232.57
91.169.34
-381.73353.53
7819
NOTE: All samples were at a protein concentration of 10 mg/mL in 20 mM HEPES pH 7.8 and incubated at 37 C
without agitation. Datasets in italics are from preseeded aggregation profiles and are excluded from average lag
time considarations.
a
Errors obtained from fitting program.
c
Errors derived using standard procedures (84) from errors in fitted and obtained from fitting program.
Figure 4.7
illustrates fits to both the primary and secondary nucleation models for pWT (A), A4S (B) and
A4T(C) and shows that pWT, A4S and A4T, like other variants of holo hSOD (77) (Y. M.
Hwang unpublished data), fit well to the secondary-nucleation equation (Equation 4.8). The fits
to the primary-nucleation equation (Equation 4.7) have a more gradual growth compared to the
fits to the secondary-nucleation equation and consistently deviate from the experimental data.
69
Figure 4.7 Holo aggregation fits to primary and secondary nucleation equations of pWT
(A), A4S (B) and A4T (C). The solid squares are experimental data points while the dotted and
dashed lines are the primary and secondary nucleation fits, respectively.
70
4.5
Discussion
4.5.1 Aggregates arise from holo hSOD dimer interface mutants A4S and A4T at
physiologically relevant pH and temperature
A protein concentration of 10 mg/mL (317 M) was used as it is within the estimated
range of hSOD concentration in motor neurons as well as a pragmatic solution for shortening the
long lag times observed during holo aggregation experiments. As mentioned in Section 4.1.2,
concentration dependence is a feature of nucleation-dependent protein aggregation, therefore,
higher protein concentrations will result in shorter lag times. Kurobe et al. and Bowling et al.
reported hSOD concentrations in erythrocyte lysate to be 0.950.07 g/mg haemoglobin and
7.01.8 g/mg of protein, respectively (108,109), which correlates to hSOD concentrations of
~10
M and 76
concentrations of 320 g/L, also assuming haemoglobin is ~total amount of protein in RBC). It
has also been reported that the concentration of hSOD in brain is ~4.5 times higher than in
erythrocytes, corresponding to hSOD concentrations of ~45-340
M (100,109). Moreover,
immunohistochemical studies indicate that motor neurons have higher concentrations of hSOD
compared to other neuronal populations (110). Although other studies on hSOD aggregation
rarely have protein concentrations above 3.15 mg/mL (100
protein aggregation studies of other proteins that use protein concentrations in the range of 10
mg/mL (111-114).
The pWT and dimer interface mutants A4S and A4T used in the holo aggregation
experiments were fully metallated as confirmed by the pyrogallol activity assay and DSC. The
full metallation status of hSOD was characterized by having a specific activity of 1800200
71
U/mg and a single endothermic peak by DSC (Figure 3.3 C and D). The correlation of this
specific activity range to a full metallation status was previously confirmed with inductively
coupled plasma atomic emission spectroscopy (ICP-AES) where holo hSOD samples with
specific activities of 1800200 U/mg had Cu2+ and Zn2+ to monomer ratios of ~1 (77). The
average specific activities for A4S and A4T are 194363 U/mg and 1736260 U/mg,
respectively; this, together with the single peak observed in DSC indicates that the mutants are
fully metallated. Additionally, if aggregation occurred from a small amount of apo/mismetalated
or misfolded protein in the sample, not only should the specific activity of the sample increase as
aggregation occurs but aggregation should not restart after the aggregates are removed at the
plateau phase, since both should affectively remove the apo/mismetalated/misfolded hSOD that
may be present initially. Previous studies on holo hSOD aggregation have shown that the
specific activity of the sample decreases with time while samples from restart experiments,
where samples at plateau phase were filtered using 20 nm filters to remove the aggregates, go
through another sigmoidal aggregation profile (Y. M. Hwang unpublished data). These indicate
that the observed aggregation profiles are not due to pre-existing apo/mismetallated or midfolded
hSOD in the sample but arise from holo hSOD.
Despite several studies reporting that holo hSOD does not aggregate at physiologically
relevant conditions (37,90), we observe that holo A4S and A4T, in addition to other holo hSOD
mutants studied in the Meiering laboratory ((77) and Y.M. Hwang unpublished data), do possess
the ability to aggregate at physiologically relevant conditions of pH 7.8, 37 C and without
agitation. This discrepancy could be due to differences in experimental conditions. Thioflavin T
(ThT) dye binding was employed in the previous studies to monitor hSOD aggregation (90,91)
while our studies use light scattering to directly measure the change in particle size in solution.
72
ThT is an amyloid specific dye that binds to the surface of the -rich quaternary structure of
amyloid fibrils (115). It is worth noting that the aggregates formed in our holo aggregation
studies as well as in ALS patients are not amyloid and do not exhibit the classical features of
amyloid, such as long unbranched fibrils, ThT binding, in the case of our aggregates (Y. M.
Hwang unpublished data), and green-gold birefringence when stained with Congo red, in the
case of protein inclusions in ALS patients (11,116). It is also important to note that dye binding
assays have an inherent shortcoming in that the addition of the dyes may affect the equilibrium
between the various protein states in the sample (98). Another difference could be the shorter
range of time the samples are monitored for or a lower frequency of measurements in the
previous studies compared to the long time courses in our experiments. Identical holo hSOD
samples, monitored concomitantly, revealed that samples that were measured more frequently
have shorter lag times compared to those that were not (Y. M. Hwang unpublished data),
suggesting that the sample measurement process promotes aggregation. To minimize any
inconsistencies from the measurement process, the frequency of measurements was done in a
consistent manner (described in Section 4.2.3) for all samples.
4.5.2 Holo hSOD dimer interface mutants A4S and A4T have shorter lag times
compared to pWT
Dimer interface mutants A4S and A4T have shorter averaged lag times compared to
pWT. This is in agreement with previous data obtained in the Meiering laboratory where dimer
interface mutants A4V and I149T also have shorter lag times compared to pWT (77) (Y.M.
Hwang unpublished data). However, non-dimer interface mutants G93A/D/R/S/V, with the
exception of G85R which has an averaged lag time of ~97 h, have average lag times similar to, if
not longer than, pWT (77,91) (Y.M. Hwang unpublished data). G85R is a mutation that disrupts
73
metal binding. This implies that a disrupted dimer interface and/or metal binding increases the
aggregation propensity of holo hSOD and, perhaps, monomerization may be involved in
aggregation. Monomerization is not an uncommon step in the protein aggregation pathways of
multisubunit proteins and occurs during the aggregation of transthyretin (117) and insulin (118).
It has also been shown that the majority of hSOD mutations either decrease dimer stability or
increase dimer disscociation (36). In addition, studies using an antibody specific to the dimer
interface of hSOD have shown that aggregates in mice models and in an A4V patient contain
dimer interface exposed species (52,119). Therefore, it is reasonable to conclude that hSOD
aggregation can proceed via monomerization or partial exposure of the dimer interface.
The lag times also correlate with the thermodynamic stability of A4S and A4T. A4S has
higher thermal stability compared to A4T while both mutants are destabilized compared to pWT
(Chapter 3). Alanine 4 is located at the edge of the dimer interface and points into the core of the
-barrel. Both A4S and A4T replace a non-polar residue with a more polar residue and are also
both larger compared to alanine. Both the decrease in hydrophobicity and increase in side chain
size will likely weaken the dimer interface and destabilize the monomer. As discussed in Chapter
3, ITC of holo hSOD mutants indicate that dimer dissociation of hSOD mutants, G93A/D/R,
were only mildly weaker compared to pWT, while dimer interface mutants, A4V and I113T,
much weaker with
apo hSOD mutants also indicate that hSOD mutants, G93S, A4V, A4S, A4T, I113T, H43R
(arranged from high to low stability), have decreased monomer stability compared to pWT (79),
while chemical denaturation studies on both apo and holo hSOD mutants, G85R and E100G,
indicate that monomer stability is comparably compromised in both states (68,72,78). G85R also
has weakened metal binding. This is interesting because G85R was the only non-dimer interface
74
mutant with an appreciably shorter lag time compared to pWT (77) (Y. M. Hwang unpublished
data). The destabilization of the dimer interface and/or weakened metal binding combined with
decreased monomer stability may account for the increased aggregation propensity of dimer
interface mutants A4S, A4T, A4V and mostly likely I149T as well as G85R. The shorter lag time
for A4T compared to A4S also agrees with patient duration data where the average duration for
A4T is 1.5 (n=21) years while A4S is >3 (n=1) years (64). However, it is important to note that
correlation between stability and aggregation propensity with disease duration is not well
defined.
4.5.3 Multiple pathways of aggregation
It is worth noting that there is much variability in the aggregation profiles for each hSOD
variant in terms of lag times, final mean scattering intensities and growth rates. It is generally
accepted that protein aggregation is a complex process in which many competing reactions and
possibly different mechanisms may occur simultaneously. It is also known that the same peptide
or protein can form aggregates with different morphologies, possibly undergoing different
mechanisms/pathways, depending on differences in experimental conditions and the initial state
of the protein (98,120). An obvious possible cause and concern with regards to the variations
observed in our studies are variations between sample preparations such as metal content or
minor differences in protein/buffer concentrations. However, samples that have been split into
two and measured concurrently have produced aggregation profiles that are quite different, as
illustrated in Figure 4.6 (B). This difference in aggregation profiles of identical samples has also
been observed in previous studies (Y. M. Hwang unpublished data), suggesting that the
variations between samples are not due to differences in protein/sample quality but is intrinsic to
hSOD aggregation.
75
ALS patient as well as cell model studies reveal varying fine structures of aggregates,
from filaments of varying diameters and granule-coated fibrils to soluble aggregates (12,13). In
addition, due to the inability to differentiate between the different secondary-nucleation
mechanisms (fragmentation, branching, heterogeneous), we cannot preclude the possibility that
the variations in the aggregation profiles within each hSOD variant is due to the different
samples undergoing different combinations of aggregation pathways that lead to slightly
different aggregate distributions, thus resulting in different aggregation rates and final mean
intensities. Additionally, AFM of holo aggregation samples at the plateau phase do show
aggregates of different morphologies (Y. M. Hwang unpublished data).
76
were introduced into the recombinant pHSOD1ASlacIq vector that encodes pWT hSOD. Using
the pWT background allows quantitative thermodynamic analyses of the ALS-associated
mutants to be carried out. This is due to the lack of free cysteines in the pWT construct that
enables the reversible thermal and chemical unfolding. Seven ALS associated mutations here
were successfully introduced the pWT background via PCR-site directed mutagenesis. Those
mutants are dimer interface mutants, A4S, A4T, V148G and V148I, metal binding mutant H80R
as well as charge mutants, G37R and G41D. Of these mutants, A4S and A4T were extensively
characterized while some preliminary data was obtained for G41D and H80R.
The mutations H48R, N86D and I112T were not successfully introduced despite several
attempts to do so. However, designing new primers may increase the efficiency of the PCR-site
directed mutagenesis. I112T is a dimer interface mutation while H48R is a copper binding
mutant. On the other, like G41D, N86D is one of the few mutations that increase the net negative
charge of hSOD. These mutations, including the ones that were successfully obtained, are
structurally and chemically diverse, with disease durations ranging from 0.9 years to 14.1 years
(32,63-66). Characterization of their biophysical properties will shed more light on the
mechanisms in which ALS-associated hSOD mutants infer toxicity.
77
5.2
and I113T as well as G41D compared to pWT in the holo state. All 3 dimer interface mutant
were destabilized compared to pWT, which is similar to findings for other holo hSOD mutants
(28,71)((72) Chapter 4) ((73) Chapter 2). The mutants studied here are well fit by a dimer 2-state
unfolding model and the dimer interface mutants exhibit protein concentration dependence
characteristic of dimer 2-state unfolding at lower protein concentrations, with somewhat less
stabilization than expected at increased protein concentration, perhaps due to increased
aggregation. The dimer interface mutants had
chemical denaturation experiments of other mutants in the holo and apo states suggests that holo
A4S, A4T and I 113T may also have a weakened dimer interface.
Preliminary data on H80R indicate that the mutant has comparable stability to pWT in the
apo state. However, the as isolated H80R
indicating that either H80R does not have any bound metals or it is only slightly stabilized by
metal binding. All batches of purified H80R contained fragments of 7.2 kDa and 8.6 kDa in size.
These fragments were found to correspond to the protein being cleaved between R69 and K70.
The mechanism of cleavage is still unknown. The two most likely mechanisms are aberrant
copper chemistry, which has been observed for mutant hSOD (40,53) or
susceptibility to proteolysis.
78
an enhanced
at higher protein
concentrations (Figure 3.4). However, there are very few data points at higher protein
concentrations. To determine if the
values to A4S and A4T in DSC in the holo form, it is likely that the
s of
variant should be verified using another technique, e.g. using ICP-AES. Information from metal
analysis could provide more insight to the mechanism of H80R fragmentation as the presence of
copper is required for aberrant chemistry to occur. Depending on the cause of the protein
cleavage, a modified protein purification protocol can then be developed to obtain non-cleaved
H80R for further analysis.
79
5.3
nucleation mechanism. This is contradictory to studies that claim holo hSOD is too stable to
aggregate (89). The lag times of the mutants were quantitatively determined by fitting the
aggregation profiles to an empirical sigmoidal function. The lag times for the mutants are shorter
than for pWT with A4T having a greater decrease in lag time compared to A4S. This correlates
with the stability of the mutants determined using DSC where A4T is more thermally
destabilized than A4S, while both mutants were destabilized compared to pWT. The shorter lag
times of dimer interface mutants (A4S, A4T, A4V and I149T) compared to pWT and non-dimer
interface mutants suggests that a weakened dimer interface increases aggregation propensity.
However, aggregation experiments on various holo hSOD variants in the Meiering lab clearly
indicate that aggregation can arise from this form of the protein. Also, the ability of holo pWT to
aggregate at physiological pH and temperature and without agitation suggests a possible
connection between fALS and sALS.
5.3.1 Future work
Long term objectives of the holo aggregation studies include determining the effects of
many different types of hSOD mutations on aggregation propensity and elucidation of the
mechanism(s) by which aggregation occurs. An interesting avenue to pursue is the charge
mutations such as G37R and G41D to further investigate the effects of decreasing or increasing
the overall net charge of the protein. So far, G85R is the only mutant studied in our laboratory
that has comparable lag times to dimer interface mutants. The other non-dimer interface mutants
appear to have lag times similar to if not longer than pWT (Y. M. Hwang unpublished data).
80
Aggregation of holo hSOD can be compared to aggregation of other forms of hSOD as well as
other natively folded proteins that aggregate and cause disease. This should provide valuable
insights into fundamental aspects of aggregation mechanisms, and contribute to understanding
the toxic nature of protein aggregation in ALS and other PCDs.
81
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
Valentine JS, Doucette PA, Zittin Potter S. Copper-zinc superoxide dismutase and
amyotrophic lateral sclerosis. Annu Rev Biochem 2005;74:563-93.
Orrell RW. Amyotrophic lateral sclerosis: copper/zinc superoxide dismutase (SOD1)
gene mutations. Neuromuscul Disord 2000;10(1):63-8.
Rosen DR, Siddique T, Patterson D, Figlewicz DA, Sapp P, Hentati A, Donaldson D,
Goto J, O'Regan JP, Deng HX and others. Mutations in Cu/Zn superoxide dismutase gene
are associated with familial amyotrophic lateral sclerosis. Nature 1993;362(6415):59-62.
Pasinelli P, Brown RH. Molecular biology of amyotrophic lateral sclerosis: insights from
genetics. Nat Rev Neurosci 2006;7(9):710-23.
Boillee S, Vande Velde C, Cleveland DW. ALS: a disease of motor neurons and their
nonneuronal neighbors. Neuron 2006;52(1):39-59.
Hallewell RA, Imlay KC, Lee P, Fong NM, Gallegos C, Getzoff ED, Tainer JA, Cabelli
DE, Tekamp-Olson P, Mullenbach GT and others. Thermostabilization of recombinant
human and bovine CuZn superoxide dismutases by replacement of free cysteines.
Biochem Biophys Res Commun 1991;181(1):474-80.
Andersen PM. Genetic factors in the early diagnosis of ALS. Amyotroph Lateral Scler
Other Motor Neuron Disord 2000;1 Suppl 1:S31-42.
Figlewicz DA, Orrell RW. The genetics of motor neuron diseases. Amyotroph Lateral
Scler Other Motor Neuron Disord 2003;4(4):225-31.
Bordo D, Djinovic K, Bolognesi M. Conserved patterns in the Cu,Zn superoxide
dismutase family. J Mol Biol 1994;238(3):366-86.
Stathopulos PB, Rumfeldt JA, Karbassi F, Siddall CA, Lepock JR, Meiering EM.
Calorimetric analysis of thermodynamic stability and aggregation for apo and holo
amyotrophic lateral sclerosis-associated Gly-93 mutants of superoxide dismutase. J Biol
Chem 2006;281(10):6184-93.
Koradi R, Billeter M, Wuthrich K. MOLMOL: a program for display and analysis of
macromolecular structures. J Mol Graph 1996;14(1):51-5, 29-32.
Parge HE, Hallewell RA, Tainer JA. Atomic structures of wild-type and thermostable
mutant recombinant human Cu,Zn superoxide dismutase. Proc Natl Acad Sci U S A
1992;89(13):6109-13.
Getzoff ED, Cabelli DE, Fisher CL, Parge HE, Viezzoli MS, Banci L, Hallewell RA.
Faster superoxide dismutase mutants designed by enhancing electrostatic guidance.
Nature 1992;358(6384):347-51.
Andersen PM, Sims KB, Xin WW, Kiely R, O'Neill G, Ravits J, Pioro E, Harati Y,
Brower RD, Levine JS and others. Sixteen novel mutations in the Cu/Zn superoxide
dismutase gene in amyotrophic lateral sclerosis: a decade of discoveries, defects and
disputes. Amyotroph Lateral Scler Other Motor Neuron Disord 2003;4(2):62-73.
Andersen PM, Nilsson P, Ala-Hurula V, Keranen ML, Tarvainen I, Haltia T, Nilsson L,
Binzer M, Forsgren L, Marklund SL. Amyotrophic lateral sclerosis associated with
homozygosity for an Asp90Ala mutation in CuZn-superoxide dismutase. Nat Genet
1995;10(1):61-6.
Jonsson PA, Backstrand A, Andersen PM, Jacobsson J, Parton M, Shaw C, Swingler R,
Shaw PJ, Robberecht W, Ludolph AC and others. CuZn-superoxide dismutase in D90A
heterozygotes from recessive and dominant ALS pedigrees. Neurobiol Dis
2002;10(3):327-33.
83
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
Robberecht W, Aguirre T, Van den Bosch L, Tilkin P, Cassiman JJ, Matthijs G. D90A
heterozygosity in the SOD1 gene is associated with familial and apparently sporadic
amyotrophic lateral sclerosis. Neurology 1996;47(5):1336-9.
Khare SD, Caplow M, Dokholyan NV. FALS mutations in Cu, Zn superoxide dismutase
destabilize the dimer and increase dimer dissociation propensity: a large-scale
thermodynamic analysis. Amyloid 2006;13(4):226-35.
DiDonato M, Craig L, Huff ME, Thayer MM, Cardoso RM, Kassmann CJ, Lo TP, Bruns
CK, Powers ET, Kelly JW and others. ALS mutants of human superoxide dismutase form
fibrous aggregates via framework destabilization. J Mol Biol 2003;332(3):601-15.
Reaume AG, Elliott JL, Hoffman EK, Kowall NW, Ferrante RJ, Siwek DF, Wilcox HM,
Flood DG, Beal MF, Brown RH, Jr. and others. Motor neurons in Cu/Zn superoxide
dismutase-deficient mice develop normally but exhibit enhanced cell death after axonal
injury. Nat Genet 1996;13(1):43-7.
Bruijn LI, Houseweart MK, Kato S, Anderson KL, Anderson SD, Ohama E, Reaume AG,
Scott RW, Cleveland DW. Aggregation and motor neuron toxicity of an ALS-linked
SOD1 mutant independent from wild-type SOD1. Science 1998;281(5384):1851-4.
Goto JJ, Gralla EB, Valentine JS, Cabelli DE. Reactions of hydrogen peroxide with
familial amyotrophic lateral sclerosis mutant human copper-zinc superoxide dismutases
studied by pulse radiolysis. J Biol Chem 1998;273(46):30104-9.
Wiedau-Pazos M, Goto JJ, Rabizadeh S, Gralla EB, Roe JA, Lee MK, Valentine JS,
Bredesen DE. Altered reactivity of superoxide dismutase in familial amyotrophic lateral
sclerosis. Science 1996;271(5248):515-8.
Yim HS, Kang JH, Chock PB, Stadtman ER, Yim MB. A familial amyotrophic lateral
sclerosis-associated A4V Cu, Zn-superoxide dismutase mutant has a lower Km for
hydrogen peroxide. Correlation between clinical severity and the Km value. J Biol Chem
1997;272(14):8861-3.
Liu R, Althaus JS, Ellerbrock BR, Becker DA, Gurney ME. Enhanced oxygen radical
production in a transgenic mouse model of familial amyotrophic lateral sclerosis. Ann
Neurol 1998;44(5):763-70.
Andrus PK, Fleck TJ, Gurney ME, Hall ED. Protein oxidative damage in a transgenic
mouse model of familial amyotrophic lateral sclerosis. J Neurochem 1998;71(5):2041-8.
Beckman JS, Carson M, Smith CD, Koppenol WH. ALS, SOD and peroxynitrite. Nature
1993;364(6438):584.
Beal MF, Ferrante RJ, Browne SE, Matthews RT, Kowall NW, Brown RH, Jr. Increased
3-nitrotyrosine in both sporadic and familial amyotrophic lateral sclerosis. Ann Neurol
1997;42(4):644-54.
Ferrante RJ, Shinobu LA, Schulz JB, Matthews RT, Thomas CE, Kowall NW, Gurney
ME, Beal MF. Increased 3-nitrotyrosine and oxidative damage in mice with a human
copper/zinc superoxide dismutase mutation. Ann Neurol 1997;42(3):326-34.
Bruijn LI, Beal MF, Becher MW, Schulz JB, Wong PC, Price DL, Cleveland DW.
Elevated free nitrotyrosine levels, but not protein-bound nitrotyrosine or hydroxyl
radicals, throughout amyotrophic lateral sclerosis (ALS)-like disease implicate tyrosine
nitration as an aberrant in vivo property of one familial ALS-linked superoxide dismutase
1 mutant. Proc Natl Acad Sci U S A 1997;94(14):7606-11.
84
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
Vassall KA, Stathopulos PB, Rumfeldt JA, Lepock JR, Meiering EM. Equilibrium
thermodynamic analysis of amyotrophic lateral sclerosis-associated mutant apo Cu,Zn
superoxide dismutases. Biochemistry 2006;45(23):7366-79.
Vassall KA. Folding and stability studies on amyotrophic lateral sclerosis-associated Cu,
Zn apo superoxide dismutases [Ph. D.]. Waterloo: University of Waterloo; 2009.
Natvig DO, Imlay K, Touati D, Hallewell RA. Human copper-zinc superoxide dismutase
complements superoxide dismutase-deficient Escherichia coli mutants. J Biol Chem
1987;262(30):14697-701.
Liochev SI, Chen LL, Hallewell RA, Fridovich I. The familial amyotrophic lateral
sclerosis-associated amino acid substitutions E100G, G93A, and G93R do not influence
the rate of inactivation of copper- and zinc-containing superoxide dismutase by H2O2.
Arch Biochem Biophys 1998;352(2):237-9.
Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurement with the Folin
phenol reagent. J Biol Chem 1951;193(1):265-75.
Marklund S, Marklund G. Involvement of the superoxide anion radical in the
autoxidation of pyrogallol and a convenient assay for superoxide dismutase. Eur J
Biochem 1974;47(3):469-74.
Taylor JR. An introduction to error analysis : the study of uncertainties in physical
measurements. Mill Valley, Calif.: University Science Books; 1982. x, 270 p. p.
Liu Y, Sturtevant JM. The observed change in heat capacity accompanying the thermal
unfolding of proteins depends on the composition of the solution and on the method
employed to change the temperature of unfolding. Biochemistry 1996;35(9):3059-62.
Doucette PA. Biophysical studies of human copper-zinc superoxide dismutase and
mutants associated with the neurodegenerative disease amyotrophic lateral sclerosis
[Ph.D. ]: University of California, Los Angeles, United States; 2004.
Sandelin E, Nordlund A, Andersen PM, Marklund SSL, Oliveberg M. Amyotrophic
lateral sclerosis-associated copper/zinc superoxide dismutase mutations preferentially
reduce the repulsive charge of the proteins. Journal of Biological Chemistry
2007;282(29):21230-21236.
Calamai M, Tartaglia GG, Vendruscolo M, Chiti F, Dobson CM. Mutational analysis of
the aggregation-prone and disaggregation-prone regions of acylphosphatase. J Mol Biol
2009;387(4):965-74.
Shaw BF, Valentine JS. How do ALS-associated mutations in superoxide dismutase 1
promote aggregation of the protein? Trends Biochem Sci 2007;32(2):78-85.
Banci L, Bertini I, Durazo A, Girotto S, Gralla EB, Martinelli M, Valentine JS, Vieru M,
Whitelegge JP. Metal-free superoxide dismutase forms soluble oligomers under
physiological conditions: a possible general mechanism for familial ALS. Proc Natl Acad
Sci U S A 2007;104(27):11263-7.
Chattopadhyay M, Durazo A, Sohn SH, Strong CD, Gralla EB, Whitelegge JP, Valentine
JS. Initiation and elongation in fibrillation of ALS-linked superoxide dismutase. Proc
Natl Acad Sci U S A 2008;105(48):18663-8.
Rodriguez JA, Shaw BF, Durazo A, Sohn SH, Doucette PA, Nersissian AM, Faull KF,
Eggers DK, Tiwari A, Hayward LJ and others. Destabilization of apoprotein is
insufficient to explain Cu,Zn-superoxide dismutase-linked ALS pathogenesis. Proc Natl
Acad Sci U S A 2005;102(30):10516-21.
87
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
89